The technique of hypothermic circulatory arrest (HCA) is an established neuroprotective strategy allowing complex repairs of the thoracic aorta and congenital cardiac malformations. Despite its utility in clinical medicine, HCA is not without significant neurological sequelae, including intellectual and neuropsychomotor impairment, seizures, choreoathetosis, delayed development, and stroke. This application builds on our pioneering work delineating critical neurochemical mechanisms of excitotoxicity and neuroinflammation in a translational model of brain injury from HCA that is directly relevant to the support techniques currently used daily in patients undergoing complex heart and aortic surgery. We previously showed that valproic acid (VPA) can mitigate excitotoxic injury and that N-acetylcysteine (NAC) can attenuate neuroinflammation. However, the clinical use of VPA is limited by a severe metabolic acidosis, while the clinical use of NAC is limited by poor blood-brain barrier (BBB) penetration. We recently demonstrated that that dendrimer-drug conjugates can penetrate the BBB in this translational model of HCA and then home-in on neurons and microglia in areas of the brain that are damaged. We are thus in a unique position to evaluate this platform for targeted drug delivery to the injured brain. We hypothesize that dendrimer-drug conjugates can target delivery across the BBB selectively to injured neurons and microglia, resulting in improved efficacy at lower doses with reduced side effects, compared to systemic injections of unconjugated compounds. Our specific aims are: 1) To determine dose-response relationships for systemic administration of VPA or NAC monotherapy and for dendrimer-coupled VPA or NAC (D-VPA or D-NAC) monotherapy; 2) To assess the efficacy of combined VPA and NAC therapy on neurological injury after HCA; and 3) To assess the efficacy of targeted, combined D-VPA and D-NAC therapy on neurological injury after HCA. Evaluation of dendrimer-based therapies in a clinically relevant large-animal model will provide important information for translation to patients.